We previously compared the inhibition, insertion, and exonucleolytic removal of five currently approved antiviral nucleotide analogs on the purified human recombinant DNA polymerase gamma. The apparent Km and kcat values were determined for the incorporation of TTP, dCTP, dGTP, 2-3-dideoxy-TTP (ddTTP), 3-azido-TTP (AZT-TP), 2-3-dideoxy-CTP (ddCTP), 2-3didehydro-TTP (D4T-TP), (-)-2,3-dideoxy-3-thiacytidine (3TC-TP), and carbocyclic 2,3-didehydro-dGTP (CBV-TP). Human pol gamma readily incorporated all five analogs into DNA but with varying efficiencies. Kinetic studies indicate that the apparent in vitro hierarchy of mitochondrial toxicity for the approved NRTIs is: ddC(zalcitabine) 8805; ddI(didanosine) 8805; D4T(stavudine) > >3TC(lamivudine) >PMPA(tenofovir)> AZT(zidovudine) > CBV(abacavir). The human pol gamma utilized dideoxynucleotides and D4T-TP in vitro as efficiently as the natural deoxynucleoside triphosphates, whereas AZT-TP, 3TC-TP and CBV-TP were moderate inhibitors of chain elongation. With the exception of terminally incorporated 3TC, the pol gamma 3-5 exonuclease was inefficient at removing these five analogs from DNA and removal required enzyme levels exceeding substrate concentrations. Even though discrimination against inserting AZT and CBV makes them only moderate inhibitors in vitro, their inefficient excision suggest AZT and CBV may persist in vivo once incorporated into mtDNA by pol gamma. Finally, we found that the exonuclease activity is inhibited by AZT-monophosphate at concentrations known to occur in cells. Thus, although these analogs exert their greatest effect by insertion and chain termination of DNA synthesis, the persistence in DNA and inhibition of proofreading activity may also contribute to mitochondrial toxicity. To further investigate the potential mitochondrial toxicity from nucleoside reverse transcriptase inhibitors, we have performed studies on several new antiviral drugs. The potent antiretroviral 4'-ethynyl-2-fluoro-2'-deoxyadenosine (EFdA) is a promising experimental agent for treating HIV infection. Pre-steady-state kinetics were used to characterize the interaction of EFdA-triphosphate (EFdA-TP) with human mitochondrial DNA polymerase gamma to assess the potential for toxicity. Pol gamma incorporated EFdA-TP 4,300-fold less efficiently than dATP, with an excision rate similar to ddATP. This strongly indicates EFdA is a poor Pol gamma substrate, suggesting minimal Pol gamma-mediated toxicity, although this should be examined under clinical settings. Two novel thymidine analogs, 3'-fluoro-3'-deoxythymidine (FLT) and 2',3'-didehydro-3'-deoxy-4'-ethynylthymidine (Ed4T), have been investigated as nucleoside reverse transcriptase inhibitors (NRTIs) for treatment of HIV infection. Ed4T seems very promising in phase II clinical trials, whereas toxicity halted FLT development during this phase. To understand these different molecular mechanisms of toxicity, pre-steady-state kinetic studies were used to examine the interactions of FLT and Ed4T with wild-type (WT) human mitochondrial DNA polymerase gamma, as well as the viral target protein, WT HIV-1 reverse transcriptase (RT). We report that Ed4T-triphosphate (TP) is the first analog to be preferred over native nucleotides by RT but to experience negligible incorporation by WT pol gamma, with an ideal balance between high antiretroviral efficacy and minimal host toxicity. WT pol gamma could discriminate Ed4T-TP from dTTP 12,000-fold better than RT, with only an 8.3-fold difference in discrimination being seen for FLT-TP. A structurally related NRTI, 2',3'-didehydro-2',3'-dideoxythymidine, is the only other analog favored by RT over native nucleotides, but it exhibits only a 13-fold difference (compared with 12,000-fold for Ed4T) in discrimination between the two enzymes. We propose that the 4'-ethynyl group of Ed4T serves as an enzyme selectivity moiety, critical for discernment between RT and WT pol &#947;. We also show that the pol &#947; mutation R964C, which predisposes patients to mitochondrial toxicity when receiving 2',3'-didehydro-2',3'-dideoxythymidine to treat HIV, produced some loss of discrimination for FLT-TP and Ed4T-TP. These molecular mechanisms of analog incorporation, which are critical for understanding pol gamma-related toxicity, shed light on the unique toxicity profiles observed during clinical trials. We are continuing to utilize presteady state kinetic analysis to measure the incorporation and the selectivity of new antiviral nucleoside analogs by the human pol &#947;. Also, animal models of mitochondrial dysfunction are being pursued to evaluate the sensitivity to antivrial therapy.